Introduction
Squamous cell carcinoma of the head and neck (SCCHN) accounts for about 5% of all cancers in the West. In India, they form around 25-30% of all cancers and 60-70% of these present as locally advanced (stage III/IV) cancers.1,2 Induction chemotherapy using cisplatin based combination has yielded major response rates of upto 90% and clinical complete response rates of around 30% in locoregionally advanced head and neck SCC.3
Pre-treatment biomarkers can help to predict response to chemotherapy. The systemic inflammatory response has been regarded as an independent prognostic factor in patients with various malignancy.4 Neutrophils can facilitate tumour proliferation, invasion, and distant metastasis by secreting factors that promote tumour growth.5,6,7 Platelets are cells containing the largest quantity of growth factors, such as platelet-derived growth factor (PDGF), transforming growth factor (TGF)-β and platelet-derived endothelial cell growth factor (PD-ECGF).8 These platelet-derived growth factors are often produced in large quantities by cancer cells and contribute to cancer growth and histology. In contrast, lymphocytes, particularly cytotoxic T cells, play a crucial role in the anti - tumour immune response by promoting apoptosis and suppressing tumour growth.9,10 Accordingly, the neutrophil–lymphocyte ratio (NLR), and platelet lymphocyte ratio (PLR) a re cost-effective and simple parameters that can predict response to chemotherapy.
This study aims to find the correlation between NLR, PLR and response to induction chemotherapy in oral cavity malignancies.
Materials and Methods
Details of 32 patients with locally advanced squamous cell carcinoma of oral cavity who received induction chemotherapy from Jan 2017- March 2019 were collected and the following were recorded.
Pre - treatment total leukocyte count, neutrophil, lymphocyte, platelet counts. Using automated analyser
Tumour size - Clinical, radiological, and post operative specimen.
Three cycles of chemotherapy are given:
Premedication (Inj Hydrocortisone 100 mg, InjOndansetron 8mg, Inj Pheniramine 10mg) were given half an hour before starting chemotherapy Day 1: Paclitaxel 175 mg/m2 as 3 hrs. infusion or Docetaxel 75 mg/m2 as 2 hours infusion, Cisplatin 75 mg/m2 as divided doses on Day 1 and Day 2 as 3 hrs infusion Day 2- Day 5: 5 Fluorouracil 750 mg/m2 as 24 hrs. continuous infusion Prophylactic GM-CSF given.
21 day cycle
The pretreatment clinical measurement of maximum tumour size and post operative histopathological maximum tumour size is compared for response assessment in operated patients. The radiological (CT or MRI) change in size post three cycles of induction chemotherapy is used in non - operated patients to assess response. Patients are grouped into three categories. Those with histopathological or radiological complete response are complete responders. Patients with more than 30% of reduction in tumour size are grouped as partial responders, and those with less than 30% of reduction in tumour size are grouped as non-responders.
Blood analysis for the determination of the NLR and PLR
The blood samples were collected before the initiation of chemotherapy. Complete blood counts were measured using peripheral blood samples with automated analyser. The total count, differential neutrophil, lymphocyte count and platelet counts were measured. AbsoluteNeutrophil and lymphocyte counts calculated. Neutrophil and lymphocyte ratio obtained. Platelet count divided by absolute lymphocyte count to obtain PLR ratio.
Statistical analysis
Software used is SPSS Statistics for windows, version 23.0, Armonk, NY :IBM Corp. Released 2015
The mean value of NLR was calculated for the three groups. Kruskal Wallis test was used to calculate the significance of variation between the three groups. The significance in variation of NLR between two groups was calculated using Mann- whitney U and Wilcoxon w test.
The mean PLR was calculated and the significance in change of PLR across the three groups analysed by ANOVA test. Inter group analysis of significance in change was done by post Hoc tests.
Results for NLR
Table 1
| N | Mean | Std. Deviation | |
| Complete response group | 7 | 2.494 | .9688 |
| Partial response group | 18 | 2.622 | 1.5791 |
| Static response group | 7 | 5.614 | 2.1130 |
| Total | 32 | 3.249 | 2.0083 |
Mean NLR
The mean NLR is shown in Table 1. The mean NLR in the complete and partial response group is significantly low compared to static group.
There is a statistically significant difference in NLR when the three groups are analysed together (p=.007). [Table 2]
Table 2
| Group | N | Mean Rank | |
| NLR | Complete response group | 7 | 14.21 |
| Partial response group | 18 | 13.56 | |
| Static response group | 7 | 26.36 | |
| Total | 32 | ||
| NLR | |||
| Chi-Square | 9.931 | ||
| Df | .007 | ||
| Asymp. Sig. | 2 | ||
Kruskal wallis test
Table 3
Intergroup comparison
| (I) Group | (j) Group | P value | |
| Complete response group | Partial response group | 0.785 | |
| Partial response group | Static response group | 0.003 | |
| Static response group | Complete response group | 0.009 |
Inter group analysis revealed that there is no statistical difference in NLR between complete and partial responders (p= 0.785). There is statistical difference in the mean NLR between the partial and static responders (p=0.003). Significant difference is observed between static and complete response group. (p=0.009). [Table 3].
Table 4
ROC cutoff value for NLR
From ROC curve, a cutoff value of 3.95 is chosen.
When cutoff value of 3.95 is used 95.8% of patients with NLR below the value have either complete or partial response and 75% of patients with NLR above it are static responders [Table 4].
Results for PLR
The mean PLR in complete response group is 124. For partial response group the mean PLR is 137 and for the static response group is 180.[Table 5]
Table 5
| N | Mean | Std. Deviation | |
| Complete response group | 7 | 124.00 | 32.609 |
| Partial response group | 18 | 137.50 | 47.243 |
| Static response group | 7 | 180.43 | 44.328 |
| Total | 32 | 143.94 | 47.165 |
Mean PLR
Table 6
| Anova Test | |||||
| Sum of Squares | df | Mean Square | F | Sig. | |
| Between Groups | 12849.661 | 2 | 6424.830 | 3.320 | .050 |
| Within Groups | 56112.214 | 29 | 1934.904 | ||
| Total | 68961.875 | 31 | |||
Significance of difference between groups (PLR)
There is no statistically significant difference in the mean when all the three groups are compared (p=0.05).[Table 5].
Table 7
Inter group analysis PLR
As shown inTable 7, inter group analysis of mean PLR revealed that the re is no statistically significant difference in mean PLR between the complete and partial response groups (p=0.772). There is no significant difference in PLR between the partial and static response group (p=0.09) and static and complete response group (p=0.58).
Using ROC curves a cut-off value of 153 was determined, 94.7% of patients with PLR below 153 were responders to chemotherapy and 46.2% of patients with PLR above 153 were having static disease [Table 8]
Table 8
ROC cutoff value PLR
Combined NLR and PLR
To identify the impact of combining NLR and PLR a score of 0 is assigned for patients who have low NLR and PLR., 1 for patients who have either low NLR or PLR and 2 for patients with high PLR and NLR.
As shown in table 9, no patient with high value of either NLR or PLR was present in the complete response group. None of the patients who had a low NLR or PLR was present in the static group. In the partial response group except for 1 patient the other 16 patients had either a low NLR or PLR.
Discussion
There are different original studies and meta-analysis that show a prognostic and predictive role for NLR in solid tumours, including head-and-neck malignancies. In the present study, pre - treatment NLR and PLR have been demon strated to be correlating with response to platinum-based chemotherapy in epithelial oral cavity malignancy.
The mean NLR is significantly low in bothpartial and complete response groups compared to the patients with static response. The mean PLR is also low in responders when compared to the static group but it is not statistically significant. With a cut-off value of 3.95 for NLR and 153 for PLR, response could be predicted with high positive predictive value. When both the ratios are combined the predictive value is further increased as shown in this study.
Only a few studies have been done to evaluate the significance of pre-treatment NLR or PLR in head and neck malignancies. In a study to evaluate the association between pre- treatment NLR and outcome for locally advanced oral cavity cancers, Perisanidis et al. obtained mean NLR of 2.6 for their patients who responded to chemotherapy.11 An et al. in a similar study reported mean NLR value of 3.07 for their total cohort of patients with nasopharyngeal cancer.9 Jin et al. compared outcome and response to platinum-based chemotherapy in metastatic nasopharyngeal carcinoma. They determined a cut -off of 3.6 for NLR, based on the median of the values and showed that the response (CR + PR) rate is better with low NLR values.12 Karpathiou et al. in a retrospective study, analysed the clinical and histologic predictive factors of response to induction chemotherapy in 81 HNC patients. The patients were divided into good (62%) and poor (38%) responder groups. They selected a cut-off of 7 for NLR, and no significant difference for response rate between the two groups of low and high NLR was detected although the survival was significantly different.13 The most common primary sites in their study were hypopharynx and oropharynx (totally 86%). The discrepancy between their results and our study may be explained with the different primary sites and also the unusual NLR cut -off selected by them.
